The Wiley Network

What Comes Next? Survey Analysis and Segmentation

what-comes-next-survey-analysis-and-segmentation

January 12, 2017

 

You’ve sent out a survey, and you’ve received great responses. Hundreds, even thousands, of people answered your questions. You have looked at the results and noticed some interesting trends.

Now what?

How can you use the results of your survey towards an even greater understanding of who your respondents are?

One great way that we get deeper insight into survey respondents is through segmentation. Of course, data can be segmented in many different ways: by demographic, by geography, etc. But one of the most powerful forms of segmentation is needs-based segmentation. We sat down with Joe Stephan, president of Broadview Analytics, to talk through some of the how’s and why’s of needs-based segmentation.

Q. What is needs-based segmentation, and how can it help an organization understand their community?
A.
Segmenting target customers is a well-known business technique that is used to organize and prioritize markets. Companies have traditionally organized customers either based on how to find
them (demographics) or based on products that the company wishes to sell them (Lines of Business). A needs-based segmentation places the customers’ desires at the forefront of how a company designs and markets products or services. By understanding the markets’ needs, companies can tap into a customer’s journey and offer them the right product with the right messaging at the right time.

Q. Where do you start with this analysis? How do you get from raw data to an audience segment?
A.
A needs-based segmentation requires a quantitative survey. The choice of questions is very important and must address why an individual is interested in the product, the problems they are
trying to solve, the solutions they are using, the challenges they face and their desired outcomes. The segments are created based on the survey data using statistical algorithms that find key commonalities and difference across the individuals.

While statistical methods play a key role in segmentation, the final segments have to be psychologically consistent and realistic across all of the measures. The final segments should
be easily recognized within the audience.

Q. Can you explain a few more of the techniques you use? Factor analysis, cluster analysis?
A.
The most important thing with segmentation is to invest time and effort into writing the survey questions. Start at a high level and work down to the actual statements. This step
requires a strong understanding of the community, and should include feedback from all key stakeholders.

The first analytical step used to create the segments after data collection is Factor Analysis. This technique identifies the core trends (latent factors) within a question, which helps
mitigate the variation found within surveys. For example, instead of having 15 attributes of motivations, we have four factors that show the high-level motivations. The next step uses Cluster Analysis to group segments based on similarities and differences on the factor scores. A variety of cluster solutions are typically examined to determine the one that has the best
explanatory power.

The segments are built from the clusters by comparing where each group is statistically significantly different from the others. A detailed profile is built from the list of items, providing
context into what each segment wants, why they want it and who they are.

Q. Do you have any advice for combatting bias in segmentation analysis?
A.
Many organizations are tempted to segment their CRM (customer relationship management) data based on behavior. There are many sophisticated programs that show sales trends, but those programs are unable to provide any insight into why it happened. Segmenting based on recent behavior may provide good targets for a campaign, but it does not provide context into the optimal approach. Marketing based only on insight from behavior ultimately hampers acquisition and limits growth potential. Another pitfall is when organizations rely only on feedback
from their sales reps to develop segments. Surveys give companies a chance to hear directly from the audience about recent needs and trends. There are also potential pitfalls in the execution of the survey. Best practices are to survey the entire community and not just the customer base, have a survey that is easy for the respondent to engage and include open-ended questions to gather statements from actual voices.

Q. If an organization doesn’t have the resources to do full segmentation analysis, what scalable options might they explore?
A.
The most basic way to start to segment is to listen to your audience. Social media and product review sites are excellent ways to identify why someone is engaging with your product.
One-on-one interviews can be quite revealing. An excellent resource for this is the book Buyer Personas by Adele Revella. Organizations can also look at how competitors talk to customers. Are they trying to reach a certain segment? Do they have custom products/marketing for different buyers? Once people are aware of needs-based segmentation, they begin to see instances
of its use in product marketing and advertising.

Thanks, Joe. Segmentation analysis is time-consuming but worth it! Whether used for buyers or other stakeholders like authors or society
members, it is a useful way to ensure that we are supporting our partners as best we can.

For additional resources on survey creation and analysis, see previous posts in our series here and here.

Photo: Joe Stephan of Broadview Analysis. Image credit: Joe Stephan

 

Scholarly Innovation Video

Open Access Scholarly Communications

Related Articles